Physical AI is taking AI beyond screens and into the physical world.

By enabling machines to perceive, reason, and act autonomously, it is reshaping what robots can do – and where they can operate. In robotics, this marks a fundamental shift: from fixed, pre-programmed systems to adaptive, context-aware machines that can generalize across tasks and complex environments.

Physical AI: Taking human-robot collaboration to the next level, a new report from the Capgemini Research Institute, explores how physical AI is transforming robotics, the value it unlocks, and what it will take for organizations to scale deployments safely and effectively. Drawing on a global survey of 1,678 senior executives across 15 industries, complemented by in-depth interviews across the physical AI and robotics ecosystem, the report provides a grounded view of adoption timelines, use case maturity, and the organizational readiness required to move beyond pilots. Key findings include:

  • Physical AI is widely viewed as a game-changer across multiple dimensions. Over two-thirds of executives (67%) see physical AI as game-changing, enabling robots to interpret context, adapt in real time, and operate in unstructured environments.
  • Adoption is already underway. Nearly eight in ten organizations (79%) are engaging with physical AI today, with 65% expecting to reach scale within five years.
  • Labor pressures are accelerating demand. Labor shortages (74%) and rising labor costs (69%) are pushing organizations toward adaptive, AI-enabled robotics.
  • Value extends beyond efficiency gains. Executives expect productivity, quality, resilience, and safety improvements, while nearly four in ten anticipate new revenue opportunities and 60% see previously impractical use cases becoming viable.
  • Humanoid robots inspire conviction, but timelines remain long. While 67% believe humanoids will ultimately transform their industry, average timelines to scale extend to seven years, with technology immaturity, safety, cost, and public acceptance (cited by 62%) still key barriers.

Physical AI: Taking human-robot collaboration to the next level is intended for CTOs, CIOs, CDOs, CROs and senior executives shaping robotics and automation strategies – including technology, innovation, manufacturing, supply chain, safety, and regulatory leaders – across industrial, service, and consumer-facing environments. The report outlines five priority actions for organizations to unlock the potential of physical AI:

  • Build a clear understanding of physical AI’s current capabilities and limits
  • Start with confidence-building use cases
  • Design through form exploration and avoid defaulting to humanoids
  • Redesign workflows for human-robot collaboration
  • Scale through platform-based architectures

To succeed, these actions must be anchored in trust, supported by robust safety, governance, and human-oversight guardrails, and sustained through ongoing engagement with a fast-evolving physical AI ecosystem.

To learn more about how organizations can unlock value from physical AI in robotics, and what it takes to scale safely and effectively, download the report today.